Advanced Cardiovascular Disease Diagnosis with Machine Learning: Exploring KBest, t-SNE, Grid Search, and Ensemble Methods

Cardiovascular diseases remain a major public health concern, representing a significant cause of morbidity and mortality worldwide. It affects millions of people of all ages and ethnic origins and represents the primary factor behind mortality in numerous nations, contributing to a global aggregate of roughly 17.9 million annual deaths across the globe. With the aim of developing more accurate and effective detection approaches, and as early diagnosis of these diseases is of paramount importance for better care and management of healthcare resources, machine learning algorithms have emerged as promising assets for diagnostic analysis. This study exploited four reference databases for the classification of cardiovascular diseases. Advanced machine learning techniques, including KBest feature selection, t-SNE dimensionality reduction, hyper-parameter optimization, and ensemble methods, have been applied to enhance data quality and optimize model performance. Through the evaluation of different models, remarkable results have been achieved, with accuracy, precision, and AUC-ROC all exceeding 98%. These results reveal the success of this approach for the accurate classification of cardiovascular diseases, opening up promising prospects for better management and early detection of these potentially fatal disorders.